Image Processing Toolbox. Matlab

Transcription

1 Image Processing Toolbox Matlab 1

2 1. Introduction Matlab Platform for Image/Video Processing Image Acquisition and Sampling Some Applications Aspects of Image Processing Grayscale/RGB/Index Color Images Image Format Conversion Basics of Image Display 2

3 Matlab Platform for Image/Video Processing Matrix Laboratory Internal Data Structure and Matrix Operators are natural for Image Processing Intrinsic Matrix Operators are implemented with highly efficient Algorithms Simple access to OS file system and interactive operations over the Working Space simplifies algorithm development Ready to use Signal/Image Processing Toolboxes contain highly efficient functions which cover modern and even novel algorithms Matlab Language allows Highly Structured Object Oriented Programming Environment The Full list of Image Processing Toolbox functions help images 3

4 Image Acquisition and Sampling 4

5 Image Acquisition and Sampling 5

6 Image Acquisition and Sampling Browsing File System [Fname,Fdir]=uigetfile( *.jpg, Click on Image file ); a = imread(fname); 6

7 Image Acquisition and Sampling figure; imshow(a); 7

8 Image Acquisition and Sampling imfinfo([fdir,f1name]) ans = Filename: [1x80 char] FileModDate: '25-May :06:44' FileSize: Format: 'jpg' FormatVersion: '' Width: 2592 Height: 1944 BitDepth: 24 ColorType: 'truecolor' FormatSignature: '' NumberOfSamples: 3 CodingMethod: 'Huffman' CodingProcess: 'Sequential' Comment: {} ImageDescription: ' ' Make: 'NIKON ' Model: 'E5600 ' Orientation: 1 XResolution: 300 YResolution: 300 ResolutionUnit: 'Inch' Software: 'E5600v1.0 ' DateTime: '2010:05:25 09:06:44 ' YCbCrPositioning: 'Co-sited' DigitalCamera: [1x1 struct] disp(ans.filesize)

9 Image Acquisition and Sampling B=rgb2gray(a); figure; imshow(b); figure; mesh(b); colormap(gray); 9

10 Some Applications and Examples Medical Inspection and interpretation of images obtained from X-rays, MRI or CAT scans analysis of cell images, of chromosome karyotypes. Agriculture Satellite/aerial views of land, for example to determine how much land is being used for different purposes, or to investigate the suitability of different regions for different crops, inspection of fruit and vegetables distinguishing good and fresh produce from old. Industry Automatic inspection of items on a production line, inspection of paper samples. Law enforcements Fingerprint analysis, sharpening or de-blurring of speed-camera images. 10

11 Some Aspects of Image Processing Image enhancement sharpening or de-blurring an out of focus image, highlighting edges, improving image contrast, or brightening an image, removing noise. Image restoration removing of blur caused by linear motion, removal of optical distortions, removing periodic interference. Image segmentation finding lines, circles, or particular shapes in an image, in an aerial photograph, identifying cars, trees, buildings, or roads. Image Geometry Transformation affine transformation 11

12 Some Applications and Examples 12

13 Some Applications and Examples 13

14 Some Applications and Examples 14

15 Some Applications and Examples 15

16 Grayscale Images 16

17 RGB Images 17

18 Indexed Color Images 18

19 Indexed Color Images figure; imshow(a); [y,map]=rgb2ind(a,256); Figure; image(y); colormap(map)

20 Data Types & Image Format Conversion 20

21 Basics of Image Display impixelinfo 21

22 Basics of Image Display improfile; grid on 22

23 Basics of Image Display Image c=imread( cameraman.tif ); figure; image(c); figure; image(c); colormap(gray); n=size(unique(c)); figure; image(c); colormap(gray(n));

24 2. Point Processing Arithmetic Operations Histograms LUT Processing 24

25 b=imread( blocks.tif ); Arithmetic Operations imadd imsubstract immultiply imdivide imcomplement b1=uint8(double(b)+128); Or b1=imadd(b,128); b2=imsubstruct(b,128); 25

26 Histograms p=imread( pout.tif ); figure; imshow(p); figure; imhist(p); axis tight

27 Histograms Histogram Equalization histeq() p1=histeq(p); figure; imshow(p1); figure; imhist(p1); axis tight

28 Histograms Adaptive Histogram Equalization adapthisteq() I=imread( tire.tif ); figure; imshow(i); A=adapthisteq(I, cliplimit,0.02, Distribution, rayleigh ); figure; imshow(a); 28

29 LUT Processing p=imread( pout.tif ); figure; imshow(p); hp=imhist(p); T=integr(hp)*255/prod(size(p)); figure; imshow(uint8(t(p))); function y=integr(x) y=filter(1,[1-1],x); 29

30 3. Neighbourhood Processing Filtering in Matlab Frequencies; Low and High Pass Filters Edge sharpening Non-linear Filters 30

31 Filtering in Matlab Performing Spatial Filtering 31

32 Filtering in Matlab Performing Spatial Filtering 32

33 Filtering in Matlab Performing Spatial Filtering 33

34 Filtering in Matlab Performing Spatial Filtering 34

35 Filtering in Matlab 35

36 Filtering in Matlab Separable Filters 36

37 Frequencies; Low and High Pass Filters 37

38 Frequencies; Low and High Pass Filters 38

39 Edge sharpening Unsharp Masking The idea of unsharp masking is to subtract a scaled unsharp version of the image from the original. In practice, we can achieve this aect by subtracting a scaled blurred image from the original. 39

40 Edge sharpening Unsharp Masking 40

41 Edge sharpening Unsharp Masking 41

42 Edge sharpening Unsharp Masking 42

43 Non-linear Filters 43

44 4. The Fourier Transform The 1D/2D discrete Fourier transform Fourier transforms in Matlab Fourier transforms of images Matlab Programming Concepts Filtering in the frequency domain 44

45 The 1D/2D discrete Fourier transform 45

46 Fourier transforms in Matlab fft which takes the DFT of a vector, ifft which takes the inverse DFT of a vector, fft2 which takes the DFT of a matrix, ifft2 which takes the inverse DFT of a matrix, fftshift which shifts a transform 46

47 Fourier transforms in Matlab 47

48 Fourier transforms in Matlab 48

49 Fourier transforms of images 49

50 Fourier transforms of images Examples 50

51 Fourier transforms of images Examples 51

52 Fourier transforms of images Examples 52

53 Fourier transforms of images Examples 53

54 Matlab Programming Concepts Programming Approach SW Engineering Steps Algorithm exact Definition/Specification Program Design Coding/Editing Compilation/Run Debugging/Unit Tests Final Test Release Matlab Development Steps Algorithm Coarse Definition Interactive Algorithm evaluation/debugging/unit Test Program Assembly (diary) Final Editing Final Test/Release 54

55 Matlab Programming Concepts Programming Features Matlab traditional variable: double precision complex matrix (2D) Matlab Image/Video oriented variables: int8, uint8, frame, NaN, etc; Matlab structures: cell-arrays, functions, inline function, etc; Extended help: comment block, help, doc, lookfor, whos, which, etc; Program flow controls: for...end, if end, while end, switch end, etc; Objects: figure, plot, imshow, avifile, imaqhwinfo, etc; File access functions: fprinf, imread, read, write, fopen, fclose, etc; DDE operators: ddeinit, xlsfinfo, xlsread, xlswrite, csvread, etc; GUI editor; Build-in benchmarking: tic, toc, clock, cputime, memory, etc; Half-compiler (second run is faster!) 55

56 Matlab Programming Concepts Matlab IDE Editor 56

57 Matlab Programming Concepts Programming Stile (recommendations) Maximal use of matrix notation Avoid program flow controls (no for-loops!) Maximal use logical matrix operation Vectorized file access Standard m-file structure: function [y1,y2]=myfunc(x1,x2,x3) %myfunc brief description % continue help block %Syntax: [y1,y2]=myfunc(x1,x2,x3), where %x1 first argument %Author: (not visible by help) Function body: a=x1; %comment 57

58 Filtering in the frequency domain 58

59 Filtering in the frequency domain Ideal LPF (the same dimensions as image) Filter Application 59

60 Filtering in the frequency domain Ideal HPF (the same dimensions as image) Filter Application 60

61 Filtering in the frequency domain Filter Z-profiles 61

62 Filtering in the frequency domain Butterworth Filters 62

63 Filtering in the frequency domain Gaussian Filter 63

64 Filtering in the frequency domain LP Gaussian Filter 64

65 Filtering in the frequency domain HP Gaussian Filter 65

66 5. Image Restoration Noise Cleaning salt and pepper noise Cleaning Gaussian noise Removal of periodic noise Debluring 66

67 Salt and pepper noise Noise Gaussian noise (additive) Speckle noise (multiplicative) Periodic noise 67

68 Cleaning salt and pepper noise 68

69 Cleaning Gaussian noise Image averaging It may sometimes happen that instead of just one image corrupted with Gaussian noise, we have many different copies of it. An example is satellite imaging; if a satellite passes over the same spot many times, we will obtain many different images of the same place. 69

70 Removal of periodic noise Periodic noise may occur if the imaging equipment (the acquisition or networking hardware) is subject to electronic disturbance of a repeating nature, such as may be caused by an electric motor. 70

71 Removal of periodic noise Band reject Filtering 71

72 Removal of periodic noise Notch Filtering 72

73 Debluring The result of motion blur 73

74 Debluring To deblur the image, we need to divide its transform by the transform corresponding to the blur filter. This means that we first must create a matrix corresponding to the transform of the blur: Now we can attempt to divide by this transform. 74

75 Debluring We can constrain the division by only dividing by values which are above a certain threshold. 75

76 6. Image Segmentation Segmentation refers to the operation of partitioning an image into component parts, or into separate objects. Thresholding Edge Detection The Hough transform 76

77 Thresholding 77

78 Edge Detection Matlab Function edge Use the following techniques: Prewitt Roberts Sobel Laplacian of Gaussian ( log ) Zero-crossings of a laplacian the Marr-Hildreth method 78

79 Edge Detection I=imread('circuit.tif'); figure; imshow(i); BW1 = edge(i, log'); figure; imshow(bw1); 79

80 Edge Detection I=imread('circuit.tif'); figure; imshow(i); BW2 = edge(i, sobel'); figure; imshow(bw2); 80

81 Edge Detection I=imread('circuit.tif'); figure; imshow(i); BW3 = edge(i, 'canny'); figure; imshow(bw3); 81

82 The Hough transform If the edge points found by the edge detection methods are SPARSE, the resulting edge image may consist of individual points, rather than straight lines or curves. Thus in order to establish a boundary between the regions, it might be necessary to FIT a LINE to those points. Line definition for Hough Transform 82

83 The Hough transform RGB=imread( gantrycrane.png ); figure; imshow(rgb); BW=edge(rgb2gray(RGB), canny ); figure; imshow(bw); [H,T,R]=hough(BW, Theta,[44:0.5:46]); Counts 40 RGB size is [ ] r Theta 83

84 7. Image Geometry Transformation Affine Transform Projective Transforms Composite Transform 84

85 Affine Transform T = T=maketform( affine,[.5 0 0;.5 2 0;0 0 1]); transi = imtransform(i,t); figure, imshow(transi); I=imread( cameraman.tif ); figure; imshow(i); 85